Download Abortive rabies virus central nervous infection is induction of

Journal of NeuroVirology (2000) 6, 359 ± 372
ã 2000 Journal of NeuroVirology, Inc.
www.jneurovirol.com
Abortive rabies virus central nervous infection is
controlled by T lymphocyte local recruitment and
induction of apoptosis
Anne Galelli*,1, LeõÈla Baloul1 and Monique Lafon1
1
Unit of Neurovirology and Regeneration of Nervous System, Department of Virology, Pasteur Institute, 25 ± 28 rue du Dr
Roux, 75724 Paris Cedex 15, France
Nonfatal paralysis, induced by the attenuated Pasteur strain of rabies virus, is
characterised by local and irreversible ¯accid paralysis of the inoculated limbs.
We characterised the spread and localisation of virus in the CNS of infected
mice, determined the nature of cell injury and examined the role of the immune
response. Data indicate that infection of BALB/c mice induced paralysis in 60%
of infected mice, the others recovering without sequelae. In both groups of mice,
virus was detected in restricted sub-populations of neurons from the brain and
spinal cord, and intensity of the neuropathology correlated with levels of rabies
RNA and apoptotic infected neurons. However, apoptosis of neurons and
paralysis were not due to a direct deleterious effect of the virus, but induced by a
T-dependent immune response, as evidenced by their absence in nude mice.
Paralysed and asymptomatic mice developed a similar rabies virus-speci®c
IgG2a antibody response, thus excluding the role of any modi®cation of the
humoral immune response. In contrast, three events were critically associated
with the development of neurological symptoms: the amount of virus in the
CNS, the level of apoptosis in both infected neurons and uninfected
surrounding cells and the progressive parenchymal in®ltration of CD4+ and
CD8+ T cells at the site of infection. These data suggest that during nonfatal
rabies infection, the levels of viral replication and primary degeneration of
infected neurons by apoptosis could be responsible for the in®ltration of T
lymphocytes capable of inducing secondary degeneration of neural cells.
Journal of NeuroVirology (2000) 6, 359 ± 372.
Keywords: infectious viral immunity; neuroimmunology; T lymphocytes;
apoptosis
Introduction
Rabies virus is a highly neurotropic virus (Murphy
et al, 1973; Wunner, 1987) which induces in both
humans and animals, two forms of disease (Gamaleia, 1887; Wunner, 1987). The most frequent
manifestation of disease is an acute encephalitis
which may be associated with spastic hind limb
paralysis, thought to be a direct result of the
damaging effect of the virus on neurons. Depending
on viral pathogenicity, natural resistance of certain
mouse strains, and site of inoculation, rabies
infection can generate another form of disease,
which is also characterised by paralysis of the
limbs but is not due to a direct neuropathogenic
effect of the virus (Baer et al, 1977; Iwasaki et al,
*Correspondence: A Galelli
Received 29 October 1999; revised 24 January 2000; accepted 27
March 2000
1977). This paralysis can be either ascending and
fatal or remain local and nonfatal. Nonfatal
paralysis is induced by virus strains of low
pathogenicity and is clinically characterised by
local and irreversible ¯accid paralysis of the
inoculated limbs. Initial observations correlated
the ef®cacy of the immune response with the
induction of this form of paralysis (Iwasaki et al,
1977; Smith, 1981). Further studies using athymic
or immunosuppressed mice established the importance of a virus-induced immunopathological process involving both CD4+ and CD8+ T cells in the
development of paralysis (Smith et al, 1982;
Sugamata et al, 1992; Weiland et al, 1992). Therefore, the immune response to rabies virus infection
is thought to play a dual role, both promoting
survival of the infected animals and exacerbating
disease. However, the nature of the pathological
Immunopathology of nonfatal rabies
A Galelli et al
360
immune response remains poorly understood.
Furthermore, probably due to differences in the
observed clinical phenomena (extensor spasticity in
ascending fatal paralysis or ¯accid nonfatal paralysis), controversial results were obtained concerning
the nature and localisation of cell injury as well as
the role of T cells (Chopra et al, 1980; Iwasaki et al,
1977; Sugamata et al, 1992; Weiland et al, 1992). To
resolve these contradictions and discriminate between the deleterious effects of the immune
response and the direct neuropathogenic effect of
rabies virus, we established a model of nonfatal
paralysis. BALB/c mice were intramuscularly inoculated with the Pasteur strain of the rabies virus
(PV) which induces a non-lethal infection characterised by the development of paralysis in about
60% of the infected mice, the others recovering
without sequelae (asymptomatic). We ®rst characterised the spread and localisation of rabies virus in
the central nervous system (CNS), determined the
nature of cell injury, then examined the role of
suppressor cells on the peripheral immune response. Finally, we isolated in¯ammatory cells
in®ltrating the CNS, evaluated their CD4+ and
CD8+ T cell content and analysed their ability to
express cytokine genes.
Our results provide strong evidence that infection
of BALB/c mice with the PV strain of rabies virus
induced a localised infection resulting in apoptosis
of infected neurons and that development of
paralysis was associated with increased virus
replication and apoptosis of both infected and
non-infected cells in the CNS. However, paralysis
was not due to the direct neuropathogenic effect of
rabies virus, but correlated with the in®ltration and
persistence at the site of infection of a population of
activated mononuclear cells which contained a
majority of CD8+ T cells and expressed high levels
of mRNA for speci®c cytokines with the capacity to
exert deleterious effects on neural cells.
Results
Clinical disease of BALB/c mice infected with the
PV strain of rabies virus
Normal BALB/c mice infected i.m. with 107
infectious particles of the PV strain of rabies virus
appeared healthy for the ®rst 3 days of infection,
then all developed non-speci®c signs of illness
characterised by apathy and rough fur. From day 7
of infection, mice either recovered without sequelae
(asymptomatic) and remained healthy or developed
neurological signs of disease characterised by
tremor, ataxia, and a reduction or total loss of hind
leg motility. Before day 7, mice subsequently
paralysed were indistinguishable from mice subsequently asymptomatic. As shown in Figure 1A,
about 60% of infected normal BALB/c mice
developed ¯accid paralysis between days 7 and 11
of infection, but no mice died of infection (Figure
Journal of NeuroVirology
Figure 1 Pathology induced by infection of BALB/c mice with
the PV strain of rabies virus. Groups of eight mice were
inoculated i.m. with 107 infectious particles in both hind legs.
Paralysis (A) was de®ned as the loss of hind leg mobility.
Lethality (B) was expressed as the cumulative percentage of
mortality. +/+ BALB/c mice (* ± *), CD8+ T-depleted BALB/c
mice (* ± *), and nu/nu BALB/c mice (~ ± ~).
1B). This paralysis was restricted to the inoculated
limbs and was irreversible during an observation
period of several months. When BALB/c mice were
depleted of CD8+ T cells, they all recovered from
infection without sequelae. In contrast nu/nu
BALB/c developed cachexia, died between days
12 and 14 of infection, but never developed
neurological symptoms. These data showed that
PV infection of BALB/c mice induced a nonfatal
paralysis and con®rmed that CD4+ T cells are
required for resistance to PV rabies virus while
CD8+ T cells are involved in the development of
paralysis.
Development of paralysis is correlated with
increased but controlled viral replication in the
CNS
We ®rst characterised the spread and replication of
rabies virus in the CNS of infected BALB/c mice. To
exclude possible interference by rabies-speci®c
antibodies in the detection of infectious virus, we
evaluated the presence of rabies-speci®c genomic
RNA by semi-quantitative RT ± PCR. The same
amount of RNA was ampli®ed for 30 cycles both
for viral RNA and the housekeeping gene GAPDH,
and expression of viral RNA determined relative to
that of GAPDH mRNA. As shown in Figure 2A,
rabies-speci®c RNA expression was detected in the
spinal cord of all infected normal BALB/c mice by
day 4 and a signi®cant difference was observed by
day 7, between asymptomatic and paralysed mice.
While rabies RNA expression progressively decreased in asymptomatic mice and disappeared by
day 12, development of neurological signs was
correlated with increased expression of rabies
RNA, which remained elevated on day 11 and then
decreased signi®cantly. Rabies RNA was still
detected in the spinal cord from paralysed mice at
the end of the observation period. Furthermore,
from days 7 to 11 after infection, a correlation was
observed between the expression of viral RNA in
Immunopathology of nonfatal rabies
A Galelli et al
361
Figure 2 PCR ampli®cation of PV rabies nucleoprotein genomic
RNA extracted from the CNS of infected mice. Relative rabies
genomic RNA expression was determined as a ratio of the viral
expression to that of the housekeeping gene GAPDH and
evaluated in spinal cord (A) and brain (B) of either asympto±
) or mice developing neurological signs of
matic mice (
disease (
±
). Each point represents the arithmetic
mean+s.d. of results from 3 ± 6 mice.
the spinal cord and the intensity of neurological
symptoms. For example on day 9, relative rabies
RNA expression mean values were 0.22+0.05 for
asymptomatic mice, 0.41+0.07 for mice with
partial hind limb paralysis and 1.31+0.02 for mice
with complete hind limb paralysis. Rabies RNA
expression was also observed in the brain of all
infected mice (Figure 2B), and development of
neurological signs of disease was correlated with
earlier and increased expression of rabies RNA,
which was detected on day 7 and disappeared on
day 13. These results excluded the possibility that,
in asymptomatic mice, the virus was eliminated
before invading the CNS, and showed that paralysis
was correlated with increased viral replication.
However, this replication was limited and the virus
was progressively cleared from the CNS of paralysed mice. On day 12, rabies RNA was undetected
in spinal cord and brain from CD8 depleted BALB/c
mice. In contrast, on day 12, nude mice expressing
high levels of rabies RNA (relative rabies RNA
expression: 1.49 and 2.29 for the spinal cord and the
brain, respectively) died, but did not develop
paralysis. Therefore, difference in viral expression
may contribute to, but is not the only crucial factor
involved in the development of paralysis.
Localisation of rabies virus in the CNS
We next compared the localisation of rabies virus
antigen in the CNS of asymptomatic and paralysed
BALB/c mice. In both groups of mice, rabies virus
NC was detected in the large neurons of the spinal
cord (Figure 3A). In the brain, rabies virus NC was
restricted to scattered neurons of the brain stem and
to some Purkinje cells of the cerebellum, which
were heavily infected (Figure 3B, C). Paralysed and
asymptomatic mice differed in the number of
infected neurons. On day 7, a mean value of
76+17 infected Purkinje cells was detected in 12
cerebellar sections from paralysed mice (n=2),
whereas no infected Purkinje cells were detected
in the same number of cerebellar sections from
asymptomatic mice (n=2). The percentages of
infected neurons detected on the same number of
spinal cord sections were 3.2% for asymptomatic
mice versus 15.5% for paralysed mice. Restricted
infection in the brain explained the relatively low
viral RNA expression detected using total RNA
extracted from whole brain (Figure 2B). In contrast,
as shown in Figure 3E ± H, we observed widespread
infection of rabies virus in the brain of nu/nu
BALB/c mice.
Cell death during nonfatal rabies infection
Apoptosis of infected neurons has been implicated
in the pathogenesis of acute rabies encephalitis
(Jackson and Rossiter, 1997; Theerasurakarn and
Ubol, 1998). We therefore investigated the possible
involvement of apoptosis in the development of
nonfatal paralysis. At different times after infection,
CNS tissue sections from both paralysed and
asymptomatic mice were tested simultaneously for
apoptosis by the TUNEL method and for the
presence of rabies antigen. TUNEL positive cells
were undetected in the CNS from control mice (data
not shown). In contrast, as shown in Figure 4,
TUNEL staining was observed in CNS sections from
paralysed mice and was restricted to areas positive
for rabies antigen staining (Figure 4A ± F). TUNEL
staining was observed in Purkinje cells of the
cerebellum (Figure 4B ± F), in scattered neurons of
the brain stem and in motor neurons of the spinal
cord (data not shown). In infected Purkinje cells of
the cerebellum, apoptosis led ®rst to the disappearance of cell bodies (Figure 4D), followed by the
progressive elimination of infected dendrites (Figure 4E, F). However, only a limited number of cells
were positive both for TUNEL and rabies antigen
(Figure 4A ± C). These infected TUNEL-positive
neurons were surrounded by a signi®cant number
of cells strongly positive for TUNEL only. Cells
positive for TUNEL only were ®rst observed in areas
of the granular layer of the cerebellum close to the
cell bodies of infected Purkinje cells (Figure 4A, C),
then around infected dendrites (Figure 4D, E). In the
CNS from paralysed mice, apoptosis was observed
between days 7 and 9 in the spinal cord and from
days 8 to 10 in the brain. In asymptomatic mice,
moderate numbers of cells positive for TUNEL and
rabies antigen were also observed on days 8 and 9 in
the spinal cord (data not shown) and in the
cerebellum (Figure 4G). In contrast, despite widespread infection of their CNS, nude mice exhibited
very few TUNEL-positive cells (Figure 4H, I).
Recent reports suggested that cells containing
DNA fragmentation are not necessarily apoptotic
and that TUNEL staining alone should not be
considered as satisfactory evidence for apoptosis.
TUNEL-positive sections were therefore tested for
the presence of activated caspase-3, the central
Journal of NeuroVirology
Immunopathology of nonfatal rabies
A Galelli et al
362
Figure 3 Localization of rabies virus in the CNS 7 days after infection. Parasagittal brain sections and longitudinal spinal cord
sections were double immunostained for rabies virus NC (green) and NeuN (red) speci®c for most neurons, except Purkinje cells. In
+/+ BALB/c mice, anti-rabies virus NC staining was seen in motor neurons of the spinal cord (A), and Purkinje cells of the cerebellum
(B, C), but not in the hippocampus (D). In nu/nu BALB/c mice, rabies virus NC staining was detected in all regions of the CNS,
including neurons of the spinal cord (E), Purkinje and granular layer cells of the cerebellum (F), neurons of the cortex (G), and
hippocampal neurons (H). Photographed with a 406 objective.
effector protease speci®c to the apoptotic process
(Thornberry and Lazebnik, 1998). As shown in
Figure 4J ± L), activated caspase-3 was detected in
the cytoplasm of all infected Purkinje cells which
retained their cell body integrity, thus demonstrating the apoptotic nature of cell death.
Comparable levels of anti-rabies virus antibodies
are induced in asymptomatic and paralysed mice
Previous studies suggest that rabies virus antibodies
participate in the immunopathological process
leading to paralysis (Sugamata et al, 1992; Weiland
et al, 1992). Clearance of rabies virus is dependent
on speci®c neutralising antibodies and their isotype. Differential viral clearance observed between
asymptomatic and paralysed mice could be related
to their particular antibody response to the virus.
We therefore analysed, during the course of
infection, development of the rabies-speci®c humoral immune response in individual mice. Results
presented in Figure 5 show that asymptomatic and
paralysed BALB/c mice developed comparable
levels of total IgG anti-rabies virus-speci®c antiJournal of NeuroVirology
bodies. Analysis of isotype distribution of these
antibodies demonstrated that both types of infected
mice developed comparable levels of IgG2a antibodies but did not produce IgG1 anti-rabies virusspeci®c antibodies. Similarly, at the peak of the
response, titre of rabies virus neutralising antibodies was not signi®cantly different between
asymptomatic mice (7.1+0.4 IU/ml) and paralysed
mice (7.2+1.2 IU/ml). These results exclude the
role of a defect or a switch in the humoral response
in the immunopathological process involved in the
development of paralysis. Levels of IgG anti-rabies
virus-speci®c antibodies were also determined in
CD8+ T cell depleted mice and in nude mice on day
12 after infection. CD8-depleted mice developed a
level of IgG2a comparable to that of normal mice
(4.2+0.1 versus 4.3+0.1), whereas nude mice did
not develop rabies virus-speci®c IgG antibodies.
Absence of suppressor cells in the peripheral
lymphoid organs from paralysed mice
We then examined the possibility that, in infected
mice, development of neurological signs of disease
Immunopathology of nonfatal rabies
A Galelli et al
363
Figure 5 Asymptomatic and paralysed mice develop comparable levels of rabies virus-speci®c antibodies. Rabies-speci®c
total IgG and IgG2a from mouse sera were measured by ELISA.
Individual titres were expressed as the log10 of the highest
dilution giving a value twice as high as that observed with
normal mouse serum. Each point represents the arithmetic
mean+s.d. of results from 3 ± 6 mice.
been previously observed in lethally infected mice
(Hirai et al, 1992; Wiktor et al, 1977a). We therefore
examined, during the course of infection, the ability
of spleen cells from rabies infected BALB/c mice to
respond to an in vitro stimulation with non-speci®c
T and B mitogens. As illustrated in Figure 6, 26105
spleen cells were cultured in the presence of an
optimal concentration of Con A or LPS. The ability
of spleen cells from infected mice to proliferate
varied during the course of infection. However,
there was no signi®cant difference in spleen cell
proliferation between infected mice which developed neurological signs of disease and infected
asymptomatic BALB/c mice. Therefore, development of paralysis was unrelated to an impaired
ability of lymphoid cells to proliferate.
Figure 4 Cell death in the CNS of rabies virus infected BALB/c
mice. Double immunostaining for rabies virus NC (green) and
TUNEL positive nuclei (red) in the cerebellum of paralysed (A ±
F) or asymptomatic (G) +/+ BALB/c mice and in the spinal cord
(H) and cerebellum (I) of nu/nu BALB/c mice, after infection by
the PV stain of rabies virus. Detection of rabies virus NC (A),
TUNEL-positive nuclei (B), and co-localisation of NC and
TUNEL (yellow) in the Purkinje cell bodies from paralysed mice
(C). Disappearance of infected Purkinje cell bodies (D, arrow),
followed by the appearance of TUNEL-positive cells around
infected dendrites (E), then by the elimination of infected
dendrites and TUNEL-positive cells in the cerebellum of
paralysed mice (F). Double immunostaining for rabies virus NC
(green) and activated caspase-3 (red) in the cerebellum of
paralysed +/+ BALB/c mice (J ± L). Detection of rabies NC (J),
activated caspase-3 (K) and co-localisation of rabies virus NC
and activated caspase-3 (L) in the Purkinje cell bodies from
paralysed mice, at day 8 of infection. Photographed with a 406
(A ± I) or a 606 objective (J ± L).
was associated with impaired T cell function, due to
the induction of suppressor CD8+ T cells, as has
Cell in®ltration in the CNS of rabies infected mice
Previous histological studies on brain sections from
rabies infected mice revealed a perivascular in®ltration of CD4+ and CD8+ T lymphocytes and the
presence of CD8+ T lymphocytes throughout the
parenchyma (reviewed in; Wunner (1987) and
Sugamata et al (1992)). To further de®ne the role
of these CNS in®ltrating T cells, cells were isolated
at different time points after infection, phenotypically characterised by FACS and quanti®ed. Due to
the low numbers of in®ltrating mononuclear cells,
FACS analysis was performed on pooled cells
isolated from brain and spinal cord of 2 to 3 mice
per group. As illustrated in Figure 7, a low number
of in®ltrating CD4+ and CD8+ T lymphocytes
(600+87 and 290+65, respectively) was steadily
isolated from the CNS of uninfected control BALB/c
mice. After infection a modest but signi®cant
increase in the number of in®ltrating CD4+ and
CD8+ T cells was observed in all infected mice
(P50.02 and P50.01 for CD4+ and CD8+ T cells
respectively), peaking on day 5. The CD4 to CD8
ratio of these in®ltrating T cells was close to that
observed in the spleen or lymph nodes of these mice
(1.8 versus 2 for the spleen). From day 7 onwards, a
Journal of NeuroVirology
Immunopathology of nonfatal rabies
A Galelli et al
364
Figure 6 Spleen cells from asymptomatic and paralysed rabiesinfected BALB/c mice proliferative in response to non-speci®c
mitogens. 26105 spleen cells/well were stimulated with an
optimal concentration of either ConA (1 mg/ml) of LPS (10 mg/ml).
The proliferative response was assessed on day 3. Results are
expressed as the mean of c.p.m.+s.d. of triplicates from 3 ± 6
mice. Mean c.p.m. in unstimulated cultures: 812+393 for
asymptomatic mice and 1018+609 for paralysed mice.
by day 14 the number of CD4+ T cells had declined
whereas the number of CD8+ T cells remained
constant. Analysis on day 12, of the distribution of
T cells in brain and spinal cord indicated that T
cells present in the brain (236103 CD4+ and
30.86103 CD8+) outnumbered T cells present in
the spinal cord (12.86103 CD4+ and 18.66103
CD8+). In a previous report, in¯ammatory cell
in®ltration in the CNS had been correlated with
protection against infection (Weiland et al, 1992).
We therefore compared T cell in®ltration in the CNS
of rabies virus infected BALB/c mice with that
induced in CBA/J mice, a strain resistant to rabiesinduced infection and paralysis (Lafon and Galelli,
1996). As shown in Figure 7, the pattern of T cell
in®ltration in these mice was similar to that of
asymptomatic BALB/c mice: only a low and
transient in®ltration of both CD4+ and CD8+ T cells
was observed in all infected CBA/J mice. Together
these data demonstrate that all CBA/J and BALB/c
mice infected with the PV strain of rabies virus
developed a CNS T cell in®ltration but that the
pattern level of in®ltration was different for
paralysed mice. Development of paralysis correlated with increased and sustained in®ltration in
the CNS by both CD4+ and CD8+ T lymphocytes,
with the latter subset becoming predominant.
Figure 7 T cell in®ltration in the CNS of rabies virus-infected
mice. Mononuclear cells in®ltrating the CNS were puri®ed on a
Percoll gradient and phenotypically characterised by two-colour
staining with either anti-CD4 or anti-CD8 and anti-Mac-1 at the
indicated times after infection. Each point represents the
arithmetic mean+s.d. of staining results from 2 ± 4 experiments
performed on pooled in®ltrating cells isolated from the brain
and spinal cord from 2 ± 3 mice per group.
T lymphocytes accumulate at the site of infection
To determine the localisation of in®ltrating T cells
in the CNS, sequential frozen sections from brain
and spinal cord were analysed for viral antigen and
for CD4+ or CD8+ cells by immuno¯uorescence.
During the ®rst phase of cell in®ltration observed in
all infected mice, CD4+ and CD8+ were detected
perivascularly (data not shown). In contrast, in mice
developing neurological signs of disease, T lymphocytes accumulated at the site of infection in brain
parenchyma (Figure 8B, C) and in the spinal cord
(Figure 8D, E). CD8+ T cells appeared more abundant
than CD4+ T cells, consistent with FACS analysis of
mononuclear cells puri®ed from the CNS. Low
numbers of CD4+ and CD8+ T cells were observed
in the brain (Figure 8G, H) and spinal cord (Figure
8I, J) of asymptomatic mice. These observations
con®rmed our above mentioned results showing
that paralysis is correlated with an increased
accumulation of T cells in the CNS. Furthermore
their localisation at the site of infection explained
the relatively low level of cells obtained after
dissociation of the CNS, due to the restricted
spreading of rabies virus in the CNS of BALB/c mice.
clear difference was observed between asymptomatic and paralysed BALB/c mice (P50.005). CNS
T cell numbers declined in mice becoming asymptomatic but remained higher than in uninfected
control mice during the observation period (P50.02
and P50.01 for CD4+ and CD8+ T cells respectively).
In contrast, in mice developing neurological signs of
disease, the number of T cells continued to increase
as paralysis developed and correlated with the
severity of the disease. Furthermore, this increase
was characterised by a progressive diminution of
the CD4 to CD8 ratio (1.7 on day 8 versus 0.53 on day
14), with CD8+ T cells outnumbering CD4+ T cells.
Therefore a different pattern of in®ltration was
observed for CD4+ and CD8+ T cells. Furthermore,
Journal of NeuroVirology
Messenger RNA expression of mediators involved in
antiviral immunity by CNS in®ltrating mononuclear
cells from paralysed mice
With the aim of de®ning the role of the CNS
in®ltrating T cells, we ®rst investigated a possible
interaction between CD8+ T cells and infected
neurons. Experiments measuring anti-rabies-CTL
Immunopathology of nonfatal rabies
A Galelli et al
365
Figure 8 Localisation of in®ltrating T cells in the CNS of paralysed and asymptomatic rabies virus infected mice. Cryostat parasagittal
sections of brain (A ± C and F ± H) and longitudinal sections of spinal cord (D, E, I, J) were double immunostained for NeuN (red) and
either rabies virus NC or CD4 or CD8 (green). Detection by immuno¯uorescent staining for rabies virus NC (green) (A and F) and
in®ltrating CD4+ (B, D, G, I) and CD8+ (C, E, H, I) T cells in the cerebellum (A ± C) and the spinal cord (D, E) of paralysed mice or in
the cerebellum (G, H) and the spinal cord (I, J) of asymptomatic mice, 12 days after infection. Photographed with a 406 objective.
activity in BALB/c mice are impeded by the lack of a
susceptible target. Furthermore, in¯ammatory cells
isolated from the CNS of paralysed mice did not
proliferate in vitro in response to ConA or to
inactivated rabies virus. Therefore, in the absence
of functional assays to assess the speci®city and the
direct cytotoxic activity of these cells, we evaluated
the ability of CNS in®ltrating cells to express mRNA
speci®c for mediators essential for antiviral activity
but also capable of exerting detrimental effects on
Journal of NeuroVirology
Immunopathology of nonfatal rabies
A Galelli et al
366
neuronal function. Messenger RNA expression was
evaluated by semi-quantitative RT ± PCR, relative to
the expression of GAPDH on the same sample. As
previously indicated for FACS analysis, mRNA
cytokine expression was evaluated on pooled cells
isolated from the brain and spinal cord of 2 to 3 mice
per group. As shown in Figure 9, the expression of
two pro-in¯ammatory cytokines with antiviral
activity, IFN-g and TNF-a, were up-regulated at
different time points of maximum in®ltration but
declined at day 14. The expression of IL-2, IL-4, and
IL-10 mRNA was also evaluated, but no mRNA was
detected (data not shown). Viral antigen recognition
by CTLs leads to the release of chemokines such as
MIP-1a, MIP-1b, RANTES and IP-10 at the site of
infection and the production of these mediators has
been directly related to cytolytic function (Biddison
et al, 1997; Price et al, 1999). CNS-in®ltrating cells
from paralysed mice were therefore tested for their
ability to express these chemokine mRNAs. As
illustrated in Figure 9, high levels of mRNA of
these mediators, essential for ef®cient ampli®cation
of the immune response and able to exert deleterious effects on neural cells, were expressed by
immune cells in®ltrating the CNS of paralysed
mice.
Discussion
The present study describes a model of nonfatal
paralysis induced in BALB/c mice by the Pasteur
strain of rabies virus. In immunocompetent mice
this infection induces a local and irreversible
¯accid paralysis of the inoculated limbs in about
60% of the infected mice, the rest recovering
without sequelae. We analysed the spread of rabies
virus in the CNS, the nature of injury and the role
of T cells in this process. We demonstrated, using
CD8+ T-depleted mice and nude mice, that development of paralysis was not due to a direct
deleterious effect of virus on neurons, but linked
to an immunopathologic process involving CD8+ T
cells and apoptosis. Development of paralysis was
dependent on three linked events: the levels of
infection and apoptosis in the CNS and the
progressive parenchymal in®ltration of CD4+ and
CD8+ T cells at the site of infection. Furthermore,
the effect of these three phenomena correlated with
the intensity of neurological symptoms. In contrast,
in asymptomatic BALB/c, the absence of neurological signs of disease was associated with low levels
of infection and apoptosis and a modest and
transient in®ltration of the CNS by both CD4+ and
CD8+ T cells.
In all infected +/+ BALB/c mice, rabies virus
antigens were detected in motor neurons of the
spinal cord and in restricted sub-populations of
neurons in the brain, and the intensity of neurological symptoms correlated with levels of rabies
RNA. As previously observed with virulent strains
(Jackson and Rossiter, 1997; Theerasurakarn and
Ubol, 1998), infection by the PV attenuated strain of
rabies virus rendered neurons susceptible to
Figure 9 RT ± PCR detection of mRNAs for soluble mediators of antiviral immunity in mononuclear cells in®ltrating the CNS from
paralysed BALB/c mice. Relative expression of cytokine-speci®c mRNA was expressed as ratio of the mediator transcript to that of the
housekeeping gene GAPDH. Each point represents the value obtained with pooled cells from three mice per group. RNA extracted
uninfected BALB/c mice spleen cells was used for day 0.
Journal of NeuroVirology
Immunopathology of nonfatal rabies
A Galelli et al
367
apoptosis and the intensity of neurological symptoms CNS also correlated with the number of
apoptotic infected neurons in the CNS. However,
in contrast to virulent rabies virus activity (Theerasurakarn and Ubol, 1998), apoptosis induced by
the PV strain of rabies virus was not due to a direct
deleterious effect of the virus but to a T-dependent
immune response triggered by the virus, as shown
by the absence of apoptosis in the heavily infected
CNS of nude mice. These data con®rmed results
from Smith et al (1982), who did not observe
necrosis or neuronophagia in the CNS of immunosuppressed rabies-infected mice, despite massive
infection of their CNS.
The absence of paralysis in CD8+ T-depleted mice
and the site speci®c recruitment of a disproportionate number of CD8+ T cells at the site of
infection in the CNS of paralysed mice also
suggested that these cells are of crucial importance
in CNS injury. Several pathways may trigger a
deleterious CD8+ T cell-mediated immune response. As shown in our experiments using CD8+
T-depleted mice and also observed by Dietzschold
et al (1992), rabies virus is cleared by speci®c
antibodies without the help of the virus-speci®c
cell-mediated immune response. However virusspeci®c CD8+ T cells can be induced, contribute to,
and regulate the response to infection and its
consequences (Wiktor et al, 1977b). Because
paralysed mice developed a speci®c antibody
response identical to that of asymptomatic mice,
we can exclude the possibility that CD8+ suppressor
T cells impair the speci®c humoral antiviral
response in paralysed mice.
As previously reported (Murphy et al, 1973;
Wunner, 1987) and con®rmed in this study,
rabies virus exclusively infects neurons in the
nervous system. To explain the role of CD8+ T
cells, one hypothesis could be that, in rabies
infected mice which become paralysed, T cells
accumulate in the brain because they recognize
rabies peptides expressed at the surface of
infected neurons in association with MHC class
I molecules and that CNS injury might result from
the destruction of infected neurons by CTLs.
Neurons of uninfected mice do not express
MHC class I molecules, and there are con¯icting
reports concerning their ability to up-regulate
MHC class I expression after infection or exposure to in¯ammatory cytokines (Duguid and
Trzepacz, 1993; Joly et al, 1991; Massa et al,
1993; Momburg et al, 1986; Neumann et al, 1997;
Sethna and Lampson, 1991; Thomas et al, 1997;
Wong et al, 1984). Furthermore, even if they can
be induced to express MHC class I molecules,
neurons may be resistant to cell-mediated lysis
(Oldstone et al, 1986). We failed to detect MHC
class I expression on rabies-infected neurons,
suggesting that direct lysis of infected neurons
by speci®c CTLs can not account for the
immunopathology observed within the CNS following infection and that other viral antigen
presentation pathways may exist.
The sequence and the nature of events leading to
paralysis is unknown, but our results suggest that
the number of infected neurons dying by apoptosis
could be a crucial event. Surprisingly, apoptosis of
infected neurons was observed both in paralysed
mice and in mice recovering without sequelae. In
asymptomatic mice, apoptosis was limited and led
to the rapid elimination of infected neurons. These
results support the view that apoptosis of infected
neurons may have a physiological role in protecting
the CNS from the progression of infection and
allowing contact between virus and immune
components. They also argue in favour of the results
of Morimoto et al (1999), suggesting that neuronal
apoptosis is a host defence mechanism to limit the
infection. The events which trigger apoptosis of
infected neurons have yet to be determined. The
observation that a low number of mononuclear cells
in®ltrate the CNS of all infected immunocompetent
mice during the ®rst week of infection and the
results from Hooper et al (1998) showing that T-cell
mediated, IFN-g dependent CNS in¯ammation
collaborates with antibodies to clear virus from
the CNS, suggest that the in®ltrating cells may
induce this clearing event. However, in paralysed
mice, the number of TUNEL-positive infected
neurons was elevated and these cells were surrounded by a huge number of non-infected TUNELpositive cells which may be either apoptotic cells or
scavenger cells containing apoptotic material.
While apoptosis is regarded as the most ef®cient
mechanism to eliminate cell debris or unwanted
cells without inducing deleterious effect on surrounding tissue, cells dying by apoptosis may
trigger damaging immune responses (Ren and
Savill, 1998). Apoptotic bodies are recognised and
engulfed by scavenger phagocytes (Savill et al,
1993) and their processing may generate epitopes
recognised by speci®c CTLs (Bellone et al, 1997;
Ronchetti et al, 1999). Although this presentation
normally induces tolerance to these antigens, it can
also lead to the induction of virus-speci®c CTLs,
depending on the relative number of cells undergoing apoptosis at a given time and on the type of
scavenger phagocytes involved in their clearance
(Albert et al, 1998).
Rabies virus-infected Purkinje cells seemed
particularly vulnerable to apoptosis and the huge
amount of infected dendrites remaining in CNS
tissue after the disappearance of apoptotic cell
bodies as well as the observation that higher
numbers of infected neurons and apoptotic cells
are present in the CNS of paralysed mice, as
compared to asymptomatic mice, may signify that
the mechanism involved in the removal of these
cells is insuf®cient and lead to the accumulation of
apoptotic material. Our results showing that inJournal of NeuroVirology
Immunopathology of nonfatal rabies
A Galelli et al
368
fected neurons undergo apoptosis suggest that viral
antigens contained in apoptotic cells are accessible
to cross-priming, thus providing MHC class I
restricted Ags (Bevan, 1976; Jondal et al, 1996).
The subsequent lysis by CTLs of the APC presenting
the viral Ags could amplify the process of apoptosis
and explain the presence of non-infected TUNEL
positive cells in the vicinity of dying infected
neurons in the CNS of paralysed mice. CTL
induction by cross-priming requires CD4+ T cell
help (Bennett et al, 1997), and this may explain the
presence of these cells in the parenchyma. Inef®cient phagocytosis of cells dying by apoptosis may
also induce harmful in¯ammatory responses.
The mechanism by which T cells mediate injury
has yet to be determined, but our results show that
part of their function could be to secrete proin¯ammatory cytokines and chemokines. Cytokines
alone, or in concert with chemokines may amplify
the recruitment of T cells and macrophages to the
site of infection and tissue injury. Activation of
cerebral endothelial cells and subsequent modi®cation of the permeability of the BBB (Ludowyk et al,
1992; Sethna and Lampson, 1991) by these mediators could facilitate the recruitment of auto-reactive
T cells which recognise self antigens produced by
degeneration of the infected neurons (Wekerle,
1993). Cytokines may also act through activation
of resident neural cells. Neural cells express
receptors for chemokines, and these molecules
could induce detrimental effects on neuronal
function (Glabinski and Ransohoff, 1999). IFN-g is
a potent activator of microglia, inducing upregulation of MHC class II, phagocytosis and
production of cytokines (Merrill and Benveniste,
1996). Cytokines may also contribute to development of paralysis by directly damaging neural cells
or indirectly via activation of macrophages and
microglial cells which secrete factors toxic for
oligodendrocytes such as nitric oxide (NO) and
TNF-a (Merrill et al, 1993; Selmaj et al, 1991; Stoll
et al, 1993; Wallach, 1997; Zajicek et al, 1992).
Thus, IFN-g can directly injure oligodendrocytes
(Vartanian et al, 1995). Therefore, Cytokines and
chemokines produced by T cells could initiate the
synthesis, by both in®ltrating and resident cells of a
cascade of molecules, capable of exerting detrimental effects on neuronal activity, and resulting in
secondary degeneration of the uninfected surrounding cells.
Studies are in progress to de®ne the speci®city
and the role of T cells in®ltrating the CNS of rabiesinfected mice and to determine which of these
mechanisms are involved in the induction of
apoptosis and development of paralysis. In conclusion, our results suggest that immune-mediated
apoptosis of infected neurons and induction of
rabies-speci®c CTLs by cross-priming of viral Ags
could initiate a cascade of deleterious events in the
neighbouring cells that escape the primary injury.
Journal of NeuroVirology
Epidemiological studies suggest a role of viral
infections in the aetiology of various neurological
diseases. Therefore, this model of nonfatal paralysis
induced by peripheral infection with a neurotropic
virus may be very helpful in understanding the
mechanisms of damaging immune responses in the
CNS.
Materials and methods
Rabies virus infection of BALB/c mice, assessment
of clinical disease and preparation of cell
populations
Six-week-old +/+ BALB/c, nu/nu BALB/c and +/+
CBA/J female mice were purchased from Janvier (St.
Berthevin, France). BALB/c mice were depleted of
CD8+ T cells by 10 daily i.v. injections (beginning 5
days before infection) of 100 mg puri®ed anti-CD8
antibody (53-6.72 rat hybridoma). Depletion was
checked on spleen cells, by ¯ow cytometry analysis
(FACS), in a pilot experiment and on all infected
mice at the end of the experiment. Groups of eight
mice were inoculated intramuscularly (i.m.) with
16107 infectious particles of rabies virus Pasteur
Virus strain (PV4) in both hind legs. Neurological
symptoms were assessed by evaluating the animal's
ability to support its body weight with its hind legs.
Paralysis was de®ned as the total loss of hind leg
mobility. At different times after infection, mice
were anaesthetised with Hypnorm (Janssen, Oxford,
UK) and perfused by intracardiac injection of 50 ml
PBS. Spleen, popliteal lymph nodes, and CNS were
removed and either used immediately or stored at
7808C.
Reagents
ConA was purchased from Sigma Chemical Co. (St.
Louis, MO, USA) and LPS from Difco Laboratories
(Detroit, MI, USA). The following antibodies were
used for immuno¯uorescence staining in FACS
analysis: PE-conjugated anti-CD4 and anti-CD8
mAbs purchased from Caltag (South San Francisco,
CA, USA); FITC-conjugated anti-CD4 and antiCD11b (anti-Mac-1) mAbs. The following primary
antibodies were used for immuno¯uorescence
staining of sections: unlabelled rat anti-CD4 and
anti-CD8 mAbs (Caltag); biotinylated mouse antiMHC class I (H-2Kd) mAb (SF1-1.1; Pharmingen);
FITC-conjugated rabbit anti-rabies nucleocapsid
(NC) Ab (Sano® Diagnostics, Marnes la Coquette,
France); mouse mAb anti-calbindin-D (Sigma) to
visualise Purkinje cells, mouse mAb anti-neuronal
nuclei (NeuN; Chemicon, Temecula, CA, USA) to
visualise other neurons; goat polyclonal IgG speci®c
for the p20 subunit of CPP32 activated by cleavage
but not procaspase-3 (Santa Cruz Biotechnology,
Santa Cruz, CA, USA) to detect activated caspase-3
in apoptotic cells. Staining with primary or
secondary biotinylated antibody was followed by
FITC-streptavidin (Caltag) or Cy3-conjugated strep-
Immunopathology of nonfatal rabies
A Galelli et al
369
tavidin (Jackson Immunoresearch, Westgrove, PA,
USA). The following af®nity pure secondary antibodies were used: FITC-conjugated goat-anti rat
FITC (Caltag); rhodamine (TRITC)-conjugated goat
anti-mouse IgG1 (H+L).
RT ± PCR
Total cellular RNA was extracted from popliteal
lymph nodes or CNS mononuclear cells, using RNA
BTM (Bioprobe, Montreuil, France), according to the
manufacturer's instructions. Integrity of RNA was
tested by electrophoresis. Two mg of total RNA was
subjected to ®rst-strand cDNA synthesis in 20 ml
reaction volume containing 10 mM Tris, 50 mM
KCl, 1.5 mM MgCL2, 1 mM of each dNTPs, 20 mg
oligo(dT16), 20 U AMV reverse transcriptase (Boehringer Mannheim, Mannheim, Germany) and 50 U
RNase inhibitor, for 40 min at 428C. For detection of
genomic RNA rabies virus, reverse transcription
was performed using 1 mM rabies virus N proteinspeci®c sense primer as described (Shankar et al,
1991), instead of oligo(dT16). The resulting cDNA
was diluted 1 : 5 with distilled water and used in
PCR. PCR ampli®cations were performed in 50 ml
reaction volume containing 10 mM Tris, 50 mM
KCl, 5 mM MgCl2, 200 mM of each dNTP, 1 mM of
each speci®c primer, and 2.5 U Taq polymerase
(Perkin-Elmer, Cetus Corp., Norwalk, CT, USA). An
initial denaturation step was performed at 948C for
3 min. The cycling program consisted of a denaturation step at 948C for 45 s, primer annealing for
1 min and elongation at 728C for 1 min. In the last
cycle, the elongation step was prolonged to 10 min.
PCR reactions were performed in a GeneAmp1 PCR
System 9600 Thermal Cycler (Perkin-Elmer).
Primers for RT ± PCR
Oligonucleotide sequence of primers and probes
were synthesised by Genset (Genset SA, Paris,
France) as follows: IFN-g (GenBank accession
number K00083) 5'-AGC TCT GAG ACA ATG
AAC GC-3' (left), 5'-GGA CAA TCT CTT CCC CAC
CC-3' (right), 5'-GTG GAA GGA CTC ATG ACC ACA
GTC CATG CC-3' (probe); TNF-a (GenBank accession number M13049), 5'-GTA TGA GAT AGC AAA
TCG GC-3' (left), 5'-CTG AAC TTC GGG GTG ATC
GG-3' (right), 5'-AGG AGG GCG TTG GCG CGC TG3' (probe); MIP-1a (GenBank accession number
X53372), 5'-AAG CAG CAG CGA GTA CCA GT-3'
(left), 5'-CTC AAG CCC CTG CTC TAC AC-3' (right),
5'-AGG TGT CAT TTT CCT GAC TAA GAG AAA C3' (probe); MIP-1b (GenBank accession number
X62502), 5'-GCT CTG TGC AAA CCT AAC CC-3'
(left), 5'-AAG AAG AGG GGC AGG AAA TC-3'
(right), 5'-AGA AGC TTT GTG ATG GAT TAC TAT
GAG AC-3' (probe); IP-10 (GenBank accession
number L07417), 5'-CCT ATC CTG CCC ACG TGT
TGA G-3' (left), 5'-CGC ACC TCC ACA TAG CTT
ACA G-3' (right), 5'-ATG GAC AGC AGA GAG CCT
GT-3' (probe); RANTES (GenBank accession num-
ber U02298), 5'-CCC TCA CCA TCA TCC CTC ACT3' (left), 5'-ATT TCT TGG GTT TGC TGT GC-3'
(right), 5'-AAG AAT ACA TCA ACT ATT TGG AGA
T-3' (probe); GAPDH (GenBank accession number
M32599) 5'-ACC ACC ATG GAG AAG GCT GG-3'
(left), 5'-CTC AGT GTA GCC CAG GAT GC-3' (right),
5'-GTG GAA GGA CTC ATG ACC ACA GTC CAT
GCC-3' (probe). Forward and reverse primers were
selected from different exons. Rabies N protein
(GenBank accession number AF033905), 5'-GGA
ATT CTC CGG AAG ACT GGA CCA GCT ATG G-3'
(sense primer), 5'-AGA ATT CCC ACT CAA GCC
TAG TGA ACG G-3' (antisense primer), 5'-GGC CGG
AAC CTA TGA CAT GTT TTT CTC CCG-3' (probe)
(Shankar et al, 1991).
Analysis and semi quantitation of PCR products
In pilot experiments, designed to control the
speci®city of ampli®cation, PCR products were
run on an agarose gel stained with ethidium
bromide, gel-puri®ed and sequenced with the same
primers, using the Dye Terminator sequencing kit
and automatic sequencing (ABI Prims 377 DNA
Sequencer; Perkin-Elmer). PCR products were
quanti®ed by ELISA after hybridisation in solution
using the PCR ELISA Boehringer kit (Boehringer
Mannheim). Identical results were obtained when
PCR products were quanti®ed by densitometry
analysis of ethidium bromide-stained agarose gels,
using a video system and ImageQuaNT software
(Molecular Dynamics, Sunnyvale, CA, USA). Therefore, this method was used for semi-quantitative
determination of gene expression. The amount of
PCR products accumulated was measured during
the exponential phase of target ampli®cation, when
none of the reaction components is limiting. The
exponential phase was determined for each set of
primers, with samples containing a high starting
copy number. Ampli®cation of an endogenous
control, the housekeeping gene glyceraldehyde-3phosphate dehydrogenase (GAPDH) was performed
to standardise the amount of cDNA added to a
reaction. The relative quantitation of genomic
rabies RNA or the relative expression of cytokinespeci®c mRNA was determined by dividing the
value obtained for the target RNA by that obtained
for GAPDH on the same sample. Targets and
endogenous controls were ampli®ed in separate
tubes. The PCR to detect rabies virus RNA and
GAPDH transcripts was performed with 30 cycles of
ampli®cation. Detection of mRNA speci®c for
cytokines and chemokines was performed using
35 cycles of ampli®cation. The presence of mRNA
speci®c for IL-2, IL-4, IL-10 was also tested after 40
cycles of ampli®cation.
Immuno¯uorescence staining of tissue sections
Dissected brains and spinal cords were embedded
in tissue-Tek O.C.T. compound (Miles Elkhart, IN,
USA), frozen on powdered dry ice and stored at
Journal of NeuroVirology
Immunopathology of nonfatal rabies
A Galelli et al
370
7808C. Parasagittal brain and longitudinal spinal
cord sections (8 mm) were cut on a cryostat, airdried, and stored at 7808C until preparation for
¯uorescence staining. Sections were ®xed in 4%
paraformaldehyde for 10 min, washed three times
in PBS, permeabilised in PBS-0.25% Triton X-100
for 10 min, and incubated for 30 min with blocking
buffer (5% goat serum in PBS) to minimise background staining. Sections were then incubated with
predetermined optimal concentrations of primary
antibodies in blocking solution for 2 h at room
temperature or overnight at 48C. After washing,
sections were incubated for 30 min with either
FITC-streptavidin, FITC-, or TRITC-labelled appropriate secondary antibodies and then mounted with
glycerol containing 1,4-diazobicyclo-(2,2,2) octane
(Sigma). Activated caspase-3 was detected in a three
step method including staining with an unlabelled
primary antibody, then with an appropriate biotinylated secondary antibody and lastly with ¯uorochrome-conjugated
streptavidin.
Controls
omitting one or the other primary antibody revealed
no staining and no cross reactivity between the
stains. Reactivity was visualised with a light
microscope equipped with ¯uorescence (Leica
Optovar). Appropriate ®lters for each ¯uorochrome
and a combined ®lter for ¯uorescein and rhodamine
were used. Images were fused in one document
using the Macintosh Adobe Photoshop 3 program.
Quantitation of rabies virus infected neurons
Six similar parasagittal brain sections and longitudinal spinal cord sections were selected from
asymptomatic (n=2) and paralysed mice (n=2) to
estimate the number of infected neurons. Sections
were double stained for either rabies virus NC and
calbindin (cerebellum) or rabies virus NC and NeuN
(spinal cord). The total number of infected Purkinje
cells per section was determined. In spinal cord
sections, neurons were counted in ®ve randomly
selected ®elds per section and a mean percentage of
infected cells was calculated.
Detection of apoptosis
Apoptosis was detected either by staining with
an antibody speci®c for the activated form of
caspase-3, the central effector of apoptosis, or by
enzymatic labelling of DNA strand breaks with
the TUNEL technique (terminal deoxynucleotidyl
transferase-mediated dUTP nick end labelling
(Gavrieli et al, 1992). Tissue sections were ®xed
and incubated with blocking solution containing
Triton X-100, as described above. Sections were
then incubated for 1 h at 378C with labelling
mix containing 0.3 U/ml terminal deoxynucleotidyl transferase (TdT) and 20 mM biot-16-dUTP
in supplied buffer (Boehringer Mannheim).
Labelled nuclei were detected by Cy3-streptavidin. Negative control was performed by omitting
TdT.
Journal of NeuroVirology
Isolation of mononuclear cells from the CNS
Brain and spinal cord tissues were dissociated
between frosted glass slides. After complete dissociation by vigorous pipetting, the brain suspension was centrifuged at 2006g for 10 min to remove
cell debris and lipids. Mononuclear cells were
isolated using a discontinuous Percoll gradient as
previously described (Clatch et al, 1990). Brie¯y,
the cell pellet was re-suspended in 5 ml of 70%
Percoll (Pharmacia, Uppsala, Sweden) in RPMI
1640. This suspension was then overlaid with equal
volumes of 37% and 30% Percoll and the gradient
was centrifuged at 5006g for 15 min at 208C. The
cell layer between the 70% : 37% Percoll interface
was recovered and brain mononuclear cells were
washed twice in medium.
Fluorescence staining for FACS analysis
Cells (106 in 50 ml) were stained with predetermined
optimal concentrations of ¯uorescent reagents in
FACS buffer (PBS supplemented with 2% BSA and
0.02% sodium azide) for 15 min at 48C. Brain T cells
were characterised by double staining with either
anti-CD4 or anti-CD8 and anti-Mac-1. Phenotypic
analysis was performed with an Excalibur ¯ow
cytometer equipped with the CellQuest software
(Becton Dickinson, Mountain View, CA, USA).
Dead cells and debris were gated out, prior to data
acquisition, using side-scatter/forward-scatter scatterplots. Ten thousand events were acquired in each
run.
Detection of anti-rabies virus antibodies
Mouse sera were tested for the presence of rabiesspeci®c antibodies by ELISA as previously described (Lafon et al, 1990). Plates were coated with
0.1 mg/well of UV-inactivated ERA strain rabies
virus. Serum samples were tested in twofold serial
dilutions. The negative control consisted of pooled
normal mouse sera. Individual titres were expressed as the log10 of the highest dilution giving a
value twice as high as that observed with negative
control.
Rabies virus neutralising antibodies were quanti®ed by the rapid ¯uorescent focus inhibition test
(RFFIT), using CVS rabies virus and BSR cells
(Montano-Hirose et al, 1993). Neutralisation titres
were expressed as IU/ml using a standard rabies
neutralising human IgG.
Proliferation assays
Triplicate spleen cell samples (106/ml) were cultured for 3 days in 200 ml of RPMI 1640 medium
(Gibco Laboratories, Grand Island, NY, USA)
supplemented with 5% FCS, 2 mM L-glutamine,
561075 M 2-ME, 1 mM sodium pyruvate, 10 mM
HEPES and antibiotics. Cultures were stimulated
with the lowest concentration of mitogen giving an
optimal response: Con A (1 mg/ml) or LPS (10 mg/
ml). Cells were pulsed with 0.25 mCi/well of
Immunopathology of nonfatal rabies
A Galelli et al
[3H]TdR (DuPont de Nemours, NEN Products, Les
Ulis, France) and harvested 4 h later. Radioactivity
was determined in a liquid scintillation beta
counter (Micro-b-6 detector Plus, Wallac, Ltd,
Turku, Finland).
371
Acknowledgements
This work was supported by an institutional grant
from the Pasteur Institute. The authors gratefully
acknowledge Dr S Nicholson for her critical
review of this manuscript.
Statistics
Standard deviation (s.d.) was used as an estimate of
variance and means were compared by using
Student's t-test.
References
Albert ML, Sauter B, Bhardwaj N (1998). Dendritic cells
acquire antigen from apoptotic cells and induce class
I restricted CTLs. Nature 392: 86 ± 89.
Baer GM, Cleary WF, Diaz AM, Perl DF (1977).
Characteristics of 11 rabies virus isolates in mice:
titers and relative invasiveness of virus incubation
period of infection, and survival of mice with
sequelae. J Infect Dis 136: 336 ± 345.
Bellone M, Lezzi G, Rovere P, Galati G, Ronchetti A,
Protti MP, Davoust J, Rugarli C, Manfredi AA (1997).
Processing of engulfed apoptotic bodies yields T cell
epitopes. J Immunol 159: 5391 ± 5399.
Bennett SRM, Carbone FR, Karamalis F, Miller JFAP,
Heath WR (1997). Induction of a CD8+ cytotoxic T
lymphocyte response by cross-priming requires cognate CD4+ T cell help. J Exp Med 186: 65 ± 70.
Bevan MJ (1976). Cross-priming for a secondary cytotoxic
response to minor H antigens with H-2 congenic cells
which do not cross-react in the cytotoxic assay. J Exp
Med 143: 1283 ± 1288.
Biddison WE, Kubota R, Kawanishi T, Taub D, Cruikshank WW, Center DM, Connor EW, Utz U, Jacobson
S (1997). Human T cell leukemia virus type I (HTLV1)-speci®c CD8+ CTL clones from patients with HTLV1-associated neurologic disease secrete proin¯ammatory cytokines, chemokines, and matrix metalloproteinase. J Immunol 159: 2018 ± 2025.
Chopra JS, Banerjee AK, Murthy JM, Pal SR (1980).
Paralytic rabies: a clinicopathological study. Brain
103: 789 ± 802.
Clatch RJ, Miller SD, Metzner R, Dal Canto MC, Lipton
HL (1990). Monocytes/macrophages isolated from the
mouse central nervous system contain infectious
Theiler's murine encephalomyelitis virus (TMEV).
Virology 176: 244 ± 254.
Dietzschold B, Kao M, Zheng YM, Chen ZY, Maul G, Fu
ZF, Rupprecht CE, Koprowski H (1992). Delineation of
putative mechanisms involved in antibody-mediated
clearance of rabies virus from the central nervous
system. Proc Natl Acad Sci USA 89: 7252 ± 7256.
Duguid J, Trzepacz C (1993). Major histocompatibility
complex genes have an increased brain expression
after scrapie infection. Proc Natl Acad Sci USA 90:
114 ± 117.
Gamaleia N (1887). Etude sur la rage paralytique chez
l'homme. Ann Inst Pasteur (Paris) 1: 63 ± 83.
Gavrieli Y, Sherman Y, Ben-Sasson SA (1992). Identi®cation of programmed cell death in situ via speci®c
labeling of nuclear DNA fragmentation. J Cell Biol
119: 493 ± 501.
Glabinski AR, Ransohoff RM (1999). Chemokines and
chemokine receptors in CNS pathology. J Neurovirol
5: 3 ± 12.
Hirai K, Kawano H, Mifune K, Fujii H, Nishizono A,
Shichijo A, Mannen K (1992). Suppression of CellMediated Immunity by Street Rabies Virus infection.
Micriobiol Immunol 36: 1277 ± 1290.
Hooper DC, Morimoto K, Bette M, Weihe E, Koprowski I,
Dietzschold B (1998). Collaboration of antibody and
in¯ammation in clearance of Rabies virus from the
central nervous system. J Virol 72: 3711 ± 3719.
Iwasaki Y, Gehrard W, Clark HF (1977). Role of host
immune response in the development of either
encephalitic or paralytic disease after experimental
rabies infection in mice. Infect Immun 18: 220 ± 225.
Jackson AC, Rossiter JP (1997). Apoptosis plays an
important role in experimental rabies virus infection.
J Virol 71: 5603 ± 5607.
Joly E, Mucke L, Oldstone MBA (1991). Viral persistence
in neurons explained by lack of major histocompatibility class I expression. Science 253: 1283 ± 1285.
Jondal MR, Schirmbeck R, Reimann J (1996). MHC class
I-restricted CTL responses to exogenous antigens.
Immunity 5: 295 ± 302.
Lafon M, Edelman L, Bouvet JP, Lafage M, Montchatre E
(1990). Human monoclonal antibodies speci®c for the
rabies virus glycoprotein and N protein. J Gen Virol
71: 1689 ± 1696.
Lafon M, Galelli A (1996). Superantigen related to virus.
Springer Semin Immunopathol 17: 307 ± 318.
Ludowyk PA, Willenborg DO, Parish CR (1992). Selective
localisatioan of neuro-speci®c T lymphocytes in the
central nervous system. J Neuroimmunol 37: 237 ±
250.
Massa PT, Ozato K, McFarlin DE (1993). Cell-type
speci®c regulation of major histocompatibility complex (MHC) class I gene expression in astrocytes,
oligodendrocytes, and neurons. Glia 8: 201 ± 207.
Merrill JE, Benveniste EN (1996). Cytokines in in¯ammatory brain lesions: helpful and harmful. Trends
Neurosci 19: 331 ± 338.
Merrill JE, Ignarro LJ, Sherman MP, Melinek J, Lane TE
(1993). Microglial cell cytotoxicity of oligodendrocytes
is mediated through nitric oxyde. J Immunol 151:
2132 ± 2141.
Momburg F, Koch N, Moller P, Moldenhauer G, Butcher
GW, Hammerling GJ (1986). Differential expression of
Ia and Ia-associated invariant chain in mouse tissues
after in vivo treatment with IFN-gamma. J Immunol
136: 940 ± 948.
Journal of NeuroVirology
Immunopathology of nonfatal rabies
A Galelli et al
372
Montano-Hirose JA, Lafage M, Weber P, Badrane H,
Tordo N, Lafon M (1993). Protective activity of a
murine monoclonal antibody against European bat
lyssavirus 1 (EBL1) infection in mice. Vaccine 11:
1259 ± 1266.
Morimoto K, Hooper DG, Spitsin S, Koprowski H,
Dietzschold B (1999). Pathogenicity of different rabies
virus variants inversely correlates with apoptosis and
rabies virus glycoprotein expression in infected
primary neuron cultures. J Virol 73: 510 ± 518.
Murphy FA, Harrison AK, Winn WC, Bauer SP (1973).
Comparative pathogenesis of rabies and rabies-like
viruses: infection of the central nervous system and
centrifugal spread of virus to peripheral tissues. Lab
Invest 29: 1 ± 16.
Neumann H, Schmidt H, Cavalie A, Jenne D, Wekerle H
(1997). Major histocompatibility complex (MHC) class
I gene expression in single neurons from the central
nervous system: differential regulation by interferon
(IFN)-g and tumor necrosis factor (TNF)-a. J Exp Med
185: 305 ± 316.
Oldstone MBA, Blount P, Southern PJ, Lampert PW
(1986). Cytoimmunotherapy for persistent virus infection reveals a unique clearance pattern from the
central nervous system. Nature 321: 239 ± 243.
Price DA, Klenerman P, Booth BL, Phillips RE, Sewell
AK (1999). Cytotoxic T lymphocytes, chemokines, and
antiviral immunity. Immunol Today 20: 212 ± 216.
Ren Y, Savill J (1998). The importance of being eaten.
Cell Death Differ 5: 563 ± 568.
Ronchetti A, Rovere P, Iezzi G, Galati G, Heltai S, Protti
MP, Garancini MP, Manfredi AA, Rugarli C, Bellone
M (1999). Immunogenicity of apoptotic cells in vivo:
Role of antigen load, antigen-presenting cells, and
cytokines. J Immunol 163: 130 ± 136.
Savill JS, Fadok V, Henson P, Haslett C (1993).
Phagocyte recognition of cells undergoing apoptosis.
Immunol Today 14: 131 ± 136.
Selmaj K, Raine CS, Farooq M, Norton WT, Brosnan CF
(1991). Cytokine cytotoxicity against oligodendrocytes.
Apoptosis induced by lymphotoxin. J Immunol 147:
1522 ± 1529.
Sethna MP, Lampson LA (1991). Immune modulation
within the brain: recruitment of in¯ammatory cells
and increased major histocompatibility antigen expression following intracerebral injection of interferon-g. J Neuroimmunol 34: 121 ± 132.
Shankar V, Dietzschold B, Koprowski H (1991). Direct
entry of Rabies virus into the central nervous system
without prior local replication. J Virol 65: 2736 ± 2738.
Smith JS (1981). Mouse model of abortive rabies
infection of the central nervous system. Infect Immun
31: 397 ± 308.
Journal of NeuroVirology
Smith JS, McClelland CL, Reid FL, Baer GM (1982). Dual
role of the immune response in street rabies virus
infection of mice. Infect Immun 35: 213 ± 221.
Stoll G, Jung S, Jander S, van der Meide P, Hartung H-P
(1993). Tumor necrosis factor-a in immune-mediated
demyelination and Wallerian degeneration of the rat
peripheral nervous system. J Neuroimmunol 45: 175 ±
182.
Sugamata M, Miyazawa M, Mori S, Spangrude GJ, Ewalt
LC, Lodmell DL (1992). Paralysis of street rabies virusinfected mice is dependent on T lymphocytes. J Virol
66: 1252 ± 1260.
Theerasurakarn S, Ubol S (1998). Apoptosis induction in
brain during the ®xed strain of rabies virus infection
correlates with onset and severity of illness. J
Neurovirol 4: 407 ± 414.
Thomas HE, Dutton R, Bartlett PF, Kay TWH (1997).
Interferon regulatory factor 1 is induced by interferong equally in neurons and glial cells. J Neuroimmunol
78: 132 ± 137.
Thornberry NA, Lazebnik Y (1998). Caspases: enemies
within. Science 281: 1312 ± 1316.
Vartanian T, Li Y, Zhao M, Stefansson K (1995).
Interferon-induced oligodendrocyte cell death: Implications for the pathogenesis of multiple sclerosis. Mol
Med 1: 732 ± 743.
Wallach D (1997). Cell death induction by TNF: a matter
of self control. Trends Biochem Sci 22: 107 ± 109.
Weiland F, Cox JH, Meyer S, Dahme E, Reddehase MJ
(1992). Rabies virus neuritic paralysis: Immunopathogenesis of nonfatal paralytic rabies. J Virol 66: 5096 ±
5099.
Wekerle H (1993). T-cell autoimmunity in the central
nervous system. Intervirology 35: 95 ± 100.
Wiktor TJ, Doherty PC, Koprowski H (1977b). In vitro
evidence of cell-mediated immunity after exposure of
mice to both live and inactivated rabies virus. Proc
Natl Acad Sci USA 74: 334 ± 338.
Wiktor TJ, Doherty PC, Koprowsky H (1977a). Suppression of cell-mediated immunity by street rabies virus.
J Exp Med 145: 1617 ± 1622.
Wong GH, Bartlett PF, Clark-Lewis I, Battye F, Schrader
JW (1984). Inducible expression of H-2 and Ia antigens
on brain cells. Nature 310: 688 ± 691.
Wunner WH (1987). Rabies Viruses-pathogenesis and
immunity. In: The Rhabdoviruses. Wagner RR (ed).
Plenum Press: New York, pp 361 ± 426.
Zajicek JP, Wing M, Scolding NJ, Compston DA (1992).
Interaction between oligodendrocytes and microglia: A
major role for complement and tumor necrosis factor
in oligodendrocyte adherence and killing. Brain 115:
1611 ± 1631.